Sub-Micrometer Dropwise Condensation under Superheated and

Oct 13, 2010 - Phone: (513)556-5023. Fax: (513) ... is dependent upon surface properties such as contact angle and thermodynamic conditions of the sys...
1 downloads 0 Views 4MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Sub-Micrometer Dropwise Condensation under Superheated and Rarefied Vapor Condition Sushant Anand and Sang Young Son* Mechanical Engineering, School of Dynamic Systems, University of Cincinnati, Cincinnati, Ohio 45221, United States Received July 1, 2010. Revised Manuscript Received September 20, 2010 Phase change accompanying conversion of a saturated or superheated vapor in the presence of subcooled surfaces is one of the most common occurring phenomena in nature. The mode of phase change that follows such a transformation is dependent upon surface properties such as contact angle and thermodynamic conditions of the system. In present studies, an experimental approach is used to study the physics behind droplet growth on a partially wet surface. Superheated vapor at low pressures of 4-5 Torr was condensed on subcooled silicon surface with a static contact angle of 60° in the absence of noncondensable gases, and the condensation process was monitored using environmental scanning electron microscopy (ESEM) with sub-microscopic spatial resolution. The condensation process was analyzed in the form of size growth of isolated droplets before a coalescence event ended the regime of single droplet growth. Droplet growth obtained as a function of time reveals that the rate of growth decreases as the droplet increases in size. This behavior is indicative of an overall droplet growth law existing over larger time scales for which the current observations in their brief time intervals could be fitted. A theoretical model based on kinetic theory further support the experimental observations indicating a mechanism where growth occurs by interfacial mass transport directly on condensing droplet surface. Evidence was also found that establishes the presence of sub-microscopic droplets nucleating and growing between microscopic droplets for the partially wetting case.

Introduction It is well-known that many phenomena of common occurrence such as spreading of liquids and formation of droplets or thin films are governed by surface free energy interaction between the liquid and solid surface. Phase change processes involving liquid-solid interactions such as condensation explicitly depend upon these surface energy interactions. Accurate calculation of surface energies is a tedious task that involves computation of intermolecular interactions, so contact angle serves as a good indicator of defining the dynamics of a liquid-gas-solid system. On the basis of contact angle, the liquid forms on a surface, and condensation can be described as filmwise, dropwise, or mixed condensation.1 For hydrophilic to super-hydrophilic surfaces, a thin sheet of liquid film is formed upon condensation. This film offers an additional resistance for heat transfer between the solid surface and the gas, thereby limiting the amount of heat transfer that can take place across the wall. For hydrophobic surfaces there is instead a large bare surface between condensing droplets, which gives them a heat transfer coefficient that is 2-10 times more than filmwise condensation.2 With a large nucleation site density (∼108/cm2 on smooth surfaces), it is possible to achieve high heat flux of 170-300 kW/m2 in dropwise condensation mode.2,3 The extent of advantage that dropwise condensation has to offer makes it a more desirable mode of heat transfer as compared to filmwise condensation. However, dropwise condensation is notoriously tough to be maintained. Solutions to promote dropwise condensation primarily work by changing the surface energy *To whom correspondence should be addressed. Phone: (513)556-5023. Fax: (513) 556-3390. E-mail: [email protected]. (1) Ma, X.; Rose, J. W.; Xu, D.; Lin, J.; Wang, B. Advances in dropwise condensation heat transfer: Chinese research. Chem. Eng. J. 2000, 78(2-3), 87–93. (2) Bejan, A.; Kraus, A. D., Heat Transfer Handbook; Wiley-Interscience: New Jersey, 2003. (3) Cengel, Y. A., Heat Transfer: A Practical Approach; McGraw-Hill Companies: New York, 2003.

17100 DOI: 10.1021/la102642r

of a solid surface by coating the surface with chemicals such as dioctadecyl disulfide or oleic acid, which give a large contact angle (>120°).4-6 Ion implantation techniques have also been used to promote the dropwise condensation mode.7,8 In general, dropwise condensation depends upon several factors such as substrate orientation, thermal properties of substrate material, steam velocity, surface characteristics, wall subcooling, and presence of noncondensable gases.9 These factors inherently affect several phenomenon associated with dropwise condensation such as droplet formation, size, nucleation density, drop size distribution, and heat transfer coefficient. Understanding droplet formation and droplet growth kinetics is of utmost importance to develop a sound comprehension of vapor-liquid dynamics. On the basis of previous experimental studies observing the microscopic nature of droplet formation, two theories have been hypothesized to predict droplet formation. One hypothesis, by Eucken, treats dropwise condensation as a heterogeneous nucleation process with droplet embryos forming at nucleation sites while the intermediate area remains dry.10,11 This theory postulates (4) Rose, J. W. Condensation heat transfer. Warme Stoffubertragung Z. 1999, 35 (6), 479–485. (5) Vemuri, S.; Kim, K. J. An experimental and theoretical study on the concept of dropwise condensation. Int. J. Heat Mass Transfer 2006, 49(3-4), 649–657. (6) Holden, K. M.; Wanniarachchi, A. S.; Marto, P. J.; Boone, D. H.; Rose, J. W. The use of organic coatings to promote dropwise condensation of steam. J. Heat Transfer 1987, 109(3), 768–774. (7) Zhao, Q.; Burnside, B. M., Dropwise Condensation of Steam on Ion Implanted Condenser Surfaces Heat Recovery Syst. 1994, 14(5), 525-534. (8) Bani Kananeh, A.; Rausch, M. H.; Fr€oba, A. P.; Leipertz, A. Experimental study of dropwise condensation on plasma-ion implanted stainless steel tubes. Int. J. Heat Mass Transfer 2006, 49(25-26), 5018–5026. (9) Abdelmessih, A. H.; Neumann, A. W.; Yang, S. W. The effect of surface characteristics on dropwise condensation. Lett. Heat Mass Transfer 1975, 2(4), 285–291. (10) Eucken, A. Energie-und Stoffaustausch an Grenzfl€achen. Naturwissenschaften 1937, 25(14), 209–218. (11) Carey, V. P., Liquid-Vapor Phase-Change Phenomena, 2nd ed.; Taylor & Francis: New York, 2006.

Published on Web 10/13/2010

Langmuir 2010, 26(22), 17100–17110

Anand and Son

that all of the condensation takes place on the drops, while no more than monolayer exists between the droplets. Experimental observations have established that dropwise condensation indeed begins as a nucleation process.12,13 The other hypothesis put forward by Jakob, postulates that condensation begins in a filmwise manner by initially covering the whole surface.14 This thin film ruptures after reaching a critical thickness (about 1 μm), where after droplets are formed. Meanwhile smaller droplets are drawn to adjacent droplets due to surface tension effect, while the droplets also grow simultaneously due to vapor-liquid transformation on them and coalescence. This model has also been supported by observations of others.15 Yongji et al. also noted that a thin film of condensate exists in the area between the droplets, and while the droplets may depart this thin film remains.16 Although the pathways of droplet formation are still under debate, the theory of droplet condensation as a heterogeneous nucleation process has become widely accepted. On the basis of this hypothesis, several models have been proposed to describe the heat transfer for dropwise condensation. Theoretical models based on this model match to a great extent with the experimental observations. McCormick and Baer proposed a model in 1963 assuming vapor condensation to take place on sub-microscopic areas.17 Gose.et al. proposed a model in which condensation occurs only on already-formed droplets.18 Several other models were then introduced incorporating effects of coalescence, sweeping droplets, drop-size distributions, etc.19 Griffith and Suk noted that surfaces having higher thermal conductivity have higher heat transfer rate for the same subcooling.20 They noted that the definition of heat transfer coefficient inherently makes dropwise condensation as a function of thermal conductivity of base material. Mikic, however, attributed this effect to nonuniform heat flux distribution around droplets and proposed “constriction resistance” as an important parameter affecting heat transfer.21 It was noted that dropwise condensation can be described if the drop-size distribution is known. Several attempts have been made in this regard.19,22,23 However, the experimental limitations associated with using optical microscopes do not allow studying droplets smaller than 10 μm, and as a result the dropwise distributions are not available for the sub-microscopic droplet regime. Obtaining this information is critical, as Graham and Griffith noted that at least 50% of heat transfer took place through drops that were less than 10 μm in diameter.24 Welch (12) Umur, A.; Griffith, P. Mechanism of Dropwise Condensation; Massachusetts Institute of Technology, Dept. of Mechanical Engineering: 1963. (13) McCormick, J. L.; Westwater, J. W. Nucleation sites for dropwise condensation. Chem. Eng. Sci. 1965, 20(12), 1021–1036. (14) Jakob, M., Heat Transfer in Evaporation and Condensation; University of Illinois: 1937. (15) Welch, J. F.; Westwater, J. W. Microscopic study of dropwise condensation. Intl. Dev. Heat Transfer 1961, 2, 302–309. (16) Yongji, S.; Dunqi, X.; Jifang, L.; Siexong, T. A study on the mechanism of dropwise condensation. Int. J. Heat Mass Transfer 1991, 34(11), 2827–2831. (17) McCormick, J. L.; Baer, E. On the mechanism of heat transfer in dropwise condensation. J. Colloid Sci. 1963, 18(3), 208–216. (18) Gose, E. E.; Mucciardi, A. N.; Baer, E. Model for dropwise condensation on randomly distributed sites. Int. J. Heat Mass Transfer 1967, 10(1), 15–22. (19) Rose, J. W., On the mechanism of dropwise condensation. Int. J. Heat Mass Transfer 1967, 10, (6). (20) Griffith, P.; Man Suk, L. The effect of surface thermal properties and finish on dropwise condensation. Int. J. Heat Mass Transfer 1967, 10(5), 697–707. (21) Mikic, B. B. On mechanism of dropwise condensation. Int. J. Heat Mass Transfer 1969, 12(10), 1311–1323. (22) Le Fevre, E. J.; Rose, J. W. In A Theory of Heat Transfer by Dropwise Condensation; Proceedings of the Third International Heat Transfer Conference, 1966; 1966; pp 362-375. (23) Abu-Orabi, M. Modeling of heat transfer in dropwise condensation. Int. J. Heat Mass Transfer 1998, 41(1), 81–87. (24) Graham, C.; Griffith, P. Drop size distributions and heat transfer in dropwise condensation. Int. J. Heat Mass Transfer 1973, 16(2), 337–346.

Langmuir 2010, 26(22), 17100–17110

Article

and Westwater further hypothesized that 97% of heat transfer was carried out by sub-microscopic droplets or through a bare area, while large droplets remained mostly absent in active heat transfer.15 Since dropwise condensation is a transient process where surface coverage of liquid on a substrate increases with time, studying droplet growth kinetics is vital to predict the associated transient nature of heat transfer. Different growth laws have been reported previously for describing growth of droplets with and without the coalescence. Initial condensation experiments on silanized surfaces with impinging water vapor showed the dependence as RRt1/4 for single droplet growth.25,26 This was also supported by theoretical and numerical simulations that attributed this growth to diffusion of sub-microscopic droplets under the effect of concentration gradient.27 Later analysis showed the dependence to be RRt1/3, which was subsequently verified when the experiments were carried under gradual introduction of vapor.28-31 Experiments using nonpolar liquid with very low contact angle have shown that these drops grow linearly with time.32 It is clear that our understanding on mechanism and factors affecting droplet growth during condensation is not fully established. The present work aims to show new evidence in regards to submicroscopic droplet formation and vapor-to-liquid transformation kinetics. Of special interest is studying the dynamics of vapor-liquid phase change under conditions of rarefied vapor environments. Lowering pressure increases the mean free path of the vapor molecules, hence the characteristic size of the minimum droplet size is increased. At 760 Torr and 17 °C, the mean free path (MFP) of water vapor is 0.11 μm, whereas at same temperature but at 4.2 Torr the MFP reaches 21.3 μm. Thus, if there is an effect of mean free path on interfacial phase change kinetics, it is rendered visually noticeable in these conditions. Previous research for water vapor condensation suggests that at higher pressures drop conduction is the limiting resistance, whereas at the lower pressures interfacial heat transfer gains more importance.19,33 A droplet growth model was presented by Umur based on kinetic theory wherein heat transfer through conduction was considered.11 Using the theory it was proved that interfacial heat transfer coefficient decreases by decreasing the operating pressure. Further, the steam side heat transfer coefficient was found to decrease with decreasing pressure. This was supported by experimental observations.34,35 However, many aspects of vapor-liquid phase change remain unresolved, due to limitations (25) Beysens, D.; Knobler, C. M. Growth of Breath Figures. Phys. Rev. Lett. 1986, 57(12), 1433–1436. (26) Fritter, D.; Knobler, C. M.; Beysens, D. A. Experiments and simulation of the growth of droplets on a surface (breath figures). Phys. Rev. A 1991, 43(6), 2858– 2869. (27) Rogers, T. M.; Elder, K. R.; Desai, R. C. Droplet growth and coarsening during heterogeneous vapor condensation. Phys. Rev. A 1988, 38(10), 5303–5309. (28) Briscoe, B. J.; Galvin, K. P. An experimental study of the growth of breath figures. Colloids Surf. 1991, 56(C), 263–278. (29) Beysens, D. The formation of dew. Atmos. Res. 1995, 39(1-3), 215–237. (30) Zhao, H.; Beysens, D. From Droplet Growth to Film Growth on a Heterogeneous Surface: Condensation Associated with a Wettability Gradient. Langmuir 1995, 11(2), 627–634. (31) Beysens, D. Dew nucleation and growth. C. R. Phys. 2006, 7(9-10), 1082– 1100. (32) Gokhale, S. J.; Plawsky, J. L.; Wayner, P. C., Jr. Experimental investigation of contact angle, curvature, and contact line motion in dropwise condensation and evaporation. J. Colloid Interface Sci. 2003, 259(2), 354–366. (33) Hiroaki, T.; Takaharu, T. A microscopic study of dropwise condensation. Int. J. Heat Mass Transfer 1984, 27(3), 327–335. (34) Tanner, D. W.; Pope, D.; Potter, C. J.; West, D., Heat transfer in dropwise condensation at low steam pressures in the absence and presence of noncondensable gas. Int. J. Heat Mass Transfer 1968, 11, (2), 181-190. (35) Shigeo, H.; Hiroaki, T. Dropwise condensation of steam at low pressures. Int. J. Heat Mass Transfer 1987, 30(3), 497–507.

DOI: 10.1021/la102642r

17101

Article

of experimental setups such as spatial resolution, contamination issues, presence of noncondensable gases, etc. To overcome the above-mentioned limitations, an environmental scanning electron microscope (ESEM) was used. Optical microscopes have limited spatial resolution, which can be overcome by using an electron microscope. ESEM is a relatively new development in the field of SEM operations. It combines the ability of a scanning electron microscope (SEM) to resolve objects at sub-microscopic resolution with the ability to perform dynamic experiments such as condensation and evaporation by using water vapor as an inside gas at lower pressures (2-20 Torr). The construction details of an ESEM are widely available.36,37 The availability of additional tools such as Peltier cooler, stressstrain stage, micro injector, and heater stage have extended the use of ESEM to more dynamical measurements. Although the physics behind the condensation-evaporation in ESEM has been investigated,38-40 application of ESEM toward studying fluid dynamics is a more recent phenomenon. Investigators have used ESEM to observe contact angles on fiber surfaces41-43 for viewing aggregation and development of colloidal films.44,45 Expressions have been developed to obtain contact angles from the visual images obtained through ESEM wherever a droplet can be measured.46 More recently, ESEM has been used toward studying fluid flows in carbon nanopipes.47-49 Only in recent times ESEM has been applied to study dynamical growth of droplets on different surfaces. Lauri et al.50 used ESEM to study droplet size growth on flat surfaces. Heterogeneous nucleation theory was used, and the droplet size increment was measured and presented. (36) Danilatos, G. D. Introduction to the ESEM instrument. Microsc. Res. Tech. 1993, 25. (37) Danilatos, G. D. Mechanisms of detection and imaging in the ESEM. J. Microsc. 1990, 160, 9–19. (38) Cameron, R. E.; Donald, A. M. Minimizing sample evaporation in the environmental scanning electron microscope. J. Microsc. 1994, 173, 227–237. (39) Meredith, P.; Donald, A. M. Study of ’wet’ polymer latex systems in environmental scanning electron microscopy: Some imaging considerations. J. Microsc. 1996, 181(1), 23–35. (40) Thiel, B. L.; Donald, A. M. The study of water in heterogeneous media using environmental scanning electron microscopy. J. Mol. Liq. 1999, 80(2-3), 207–230. (41) Wei, Q.; Mather, R. R.; Fotheringham, A. F.; Yang, R. D. Observation of wetting behavior of polypropylene microfibers by environmental scanning electron microscope. J. Aerosol Sci. 2002, 33(11), 1589–1593. (42) Wei, Q. F.; Mather, R. R.; Fotheringham, A. F.; Yang, R. D. ESEM study of wetting of untreated and plasma treated polypropylene fibers. J. Ind. Text. 2002, 32(1), 59–66. (43) Wei, Q.; Liu, Y.; Wang, X.; Huang, F. Dynamic studies of polypropylene nonwovens in environmental scanning electron microscope. Polym. Test. 2007, 26(1), 2–8. (44) Donald, A. M.; He, C.; Royall, C. P.; Sferrazza, M.; Stelmashenko, N. A.; Thiel, B. L. Applications of environmental scanning electron microscopy to colloidal aggregation and film formation. Colloids Surf., A 2000, 174(1-2), 37–53. (45) Donald, A. M. The use of environmental scanning electron microscopy for imaging wet and insulating materials. Nat. Mater. 2003, 2(8), 511–516. (46) Brugnara, M.; Della Volpe, C.; Siboni, S.; Zeni, D. Contact angle analysis on polymethylmethacrylate and commercial wax by using an environmental scanning electron microscope. Scanning 2006, 28(5), 267–273. (47) Pi?a Rossi, M.; Ye, H.; Gogotsi, Y.; Babu, S.; Ndungu, P.; Bradley, J. C. Environmental scanning electron microscopy study of water in carbon nanopipes. Nano Lett. 2004, 4(5), 989–993. (48) Rossi, M. P.; Gogotsi, Y. In situ ESEM study of liquid transport through carbon nanopipes. Microsc. Microanal. 2004, 10(SUPPL. 2), 1054–1055. (49) Yarin, A. L.; Yazicioglu, A. G.; Megaridis, C. M.; Rossi, M. P.; Gogotsi, Y. Theoretical and experimental investigation of aqueous liquids contained in carbon nanotubes. J. Appl. Phys. 2005, 97, 124309. (50) Lauri, A.; Riipinen, I.; Ketoja, J. A.; Vehkam€aki, H.; Kulmala, M. Theoretical and experimental study on phase transitions and mass fluxes of supersaturated water vapor onto different insoluble flat surfaces. Langmuir 2006, 22 (24), 10061–10065. (51) Bhushan, B.; Jung, Y. C., Wetting, adhesion and friction of superhydrophobic and hydrophilic leaves and fabricated micro/nanopatterned surfaces. J. Phys.: Condens. Matter 2008, 20, (22). (52) Jung, Y. C.; Bhushan, B. Wetting behaviour during evaporation and condensation of water microdroplets on superhydrophobic patterned surfaces. J. Microsc. 2008, 229(1), 127–140.

17102 DOI: 10.1021/la102642r

Anand and Son

Figure 1. Experimental setup for condensation observation in ESEM.

Bhushan et al. have used ESEM to observe transition effects of condensation and evaporation on patterned surfaces.51-53 ESEM images were analyzed, and droplet size measurements were obtained from the analysis to present size growth during condensation and droplet diameter decrease during evaporation. Although the lateral resolution to image samples is higher than normal optical microscopes, the imaging technique of ESEM gives a poor time resolution on the order of seconds (1-10 s). ESEM operation requires the chamber to be completely vented before introducing gas such as water vapor inside it. This effectively eliminates the probability of any noncondensable gas present in the chamber. This is of vital importance as experimental studies have shown that even a small quantity of noncondensable gases results in significant deterioration in heat transfer performance by obstructing contact between vapor and the solid surface.54 The vapor inside the chamber is maintained at a particular pressure that is user controlled with the help of a computer. Further, the vapor inside carries negligible momentum, hence samples can be imaged without any surface shear effect due to vapor velocity. The rarefied environment inside the ESEM enhances the role of interfacial forces in vapor to liquid conversion. By using this technique, sub-microscopic droplet formation and growth on a partially wetting Silicon surface was observed at much higher spatial resolution than reported elsewhere. Our primary aim is to study growth kinetics of a single droplet in a coalescence free regime, although we also present experimental results of droplet growth behavior in which coalescence and new intermittent droplet formation is also included. The experimental observations are supported by the droplet growth model of Umur and the implications of this are discussed in detail.

Experimental Setup and Methodology The experimental setup used for the present studies is illustrated in Figure 1. A bare silicon wafer was used as condensation surface. A small piece (2 mm  2 mm) of silicon wafer was prepared and cleaned thoroughly using RCA cleaning process to remove any organic or inorganic contaminants present on the surface. Just before the experiment, the silicon wafer was dipped in buffered oxide etched solution for 30 s to remove any naturally formed SiO2 layer on the wafer surface. Through this method, (53) Zheng, Y.; Han, D.; Zhai, J.; Jiang, L., In situ investigation on dynamic suspending of microdroplet on lotus leaf and gradient of wettable micro- and nanostructure from water condensation. Appl. Phys. Lett. 2008, 92, (8). (54) Le Fevre, E. J.; Rose, J. W. An experimental study of heat transfer by dropwise condensation. Int. J. Heat Mass Transfer 1965, 8(8), 1117–1133.

Langmuir 2010, 26(22), 17100–17110

Anand and Son

Article

Figure 3. Copper stage for observing droplet condensation on an inclined surface. Figure 2. Saturation vapor pressure-temperature chart for ESEM.

contact angles of around 62 ( 2° were obtained for water on the silicon surface. A K-type thermocouple was planted on the Silicon surface using thermal epoxy to measure surface temperature. Due care was taken during application of the epoxy to the surface to prevent addition of large thermal mass while at same time maintaining proper contact between thermocouple tip and surface. Experiments were carried out inside a Philips XL-30 ESEM. It is equipped with a vacuum pump and a vapor supply line that is linked with a glass bottle filled with DI water. At first, the ESEM chamber is brought to a pressure of approximately equal to 10-6 Torr by the application of the vacuum pump. Then the vapor supply line is opened. Because of the vacuum pressure inside the chamber and the supply line, the DI water in the glass bottle boils at room temperature and the vapor floods the ESEM chamber. The amount of vapor inside the chamber is computer controlled such that user designated pressure is maintained inside the chamber without significant fluctuations. For surface cooling, a Peltier cooler stage was mounted inside the ESEM. Because the saturation vapor pressure is related to saturation temperature, a dynamical process such as water condensation can be studied in ESEM by using the Peltier cooler in conjunction with controlling the pressure and consequently the relative humidity (RH) in the ESEM chamber. When RH > 100%, condensation can be observed immediately. Thermodynamically, the saturation vapor pressure and temperature dependence for ESEM is shown in Figure 2. The saturation vapor pressure-temperature curve of ESEM differs from the standard curve because the temperature of the vapor Tv above the sample is different from the sample temperature Ts. The thermodynamic relation between the saturation vapor pressure and temperature for condensation to take place can be expressed by eq 1, as given by Cameron and Donald,55  0:5 Ts Pw ¼ Psat ðTs Þ ð1Þ Tv where, Psat(Ts) is saturation vapor pressure at temperature Ts. Because the vapor inside the ESEM chamber had originated from the boiling of water under low pressure, and the water temperature was maintained at ambient, the vapor temperature is the same as the water temperature in the glass bottle (around 17 °C). (55) Cameron, R. E.; Donald, A. M. Minimizing sample evaporation in the environmental scanning electron microscope. J. Microsc. 1994, 173, 227–237.

Langmuir 2010, 26(22), 17100–17110

Thus, with respect to the pressure inside the chamber (around 4.2 Torr), the vapor is in a superheated condition. The Philips XL30 ESEM has a K-type thermocouple connector available that was used in our studies. It was connected with the K-type thermocouple attached on the Si surface and externally connected to a data acquisition device (Agilent 34970A). A LabView program was used to monitor the thermocouple temperature and save the data on a computer. To initially test the peltier readout, it was attached with the peltier surface. Substantial temperature difference between the thermocouple and temperature shown by the peltier cooler was observed. As a result it was decided to use the K-type thermocouple as the basis for calculating the saturation vapor pressure-temperature curve. According to the theoretical predictions, the saturation temperature at 4.2 Torr is -0.675 °C. However, using the K-type thermocouple, the temperature at which condensation occurred at 4.2 Torr of vapor pressure was found to be -0.5 °C. In absence of proper calibration of the thermocouple and the inherent uncertainty in K-type thermocouple (( 0.5 °C), this degree of temperature difference can be anticipated in the measurements. Droplet growth was visualized in two ways with respect to substrate plane—from the top and from an almost vertical stance. For observing the substrate on the top, it was attached to the peltier surface using thermally conductive carbon tape. A special copper stage was prepared to rest on the peltier cooler to observe growth of droplets from side view so as to observe contact angle changes. The copper stage had an inclined surface of 80° as shown in Figure 3. The bare silicon wafer was placed on the inclined surface using thermally conductive carbon tape. In this arrangement, the thermocouple was attached to the copper stage instead of the silicon surface to allow proper viewing. The imaging of specimens inside the chamber was done using a gaseous secondary electron detector (GSED). The ESEM imaging involves an electron beam scanning across a given area. Electron beam parameters such as dwell time and scan rate determine the amount of time taken to scan a given area visible on screen. These parameters were adjusted, and a small area was selected to give a scan time of 2-10 s with adequate clarity. The imaging detector feeds live images to the computer. A BNC video output was used to divert the image signals to a Sony DVCAM, and live video recording was done. The timing of the DVCAM was manually synchronized with a computer to match temperature conditions with associated imaging signal. Once the system was set up, the first step involved venting the chamber and filling it with pure water vapor. An eight cycle pumpdown sequence between 0.5 and 9 Torr was performed to make sure air is properly vented out and that the chamber is filled DOI: 10.1021/la102642r

17103

Article

Anand and Son

with saturated vapor pressure at the end.55 The presence of a foreign surface in the vicinity of metastable vapor lowers the energy barrier required to cause nucleation, as compared to homogeneous nucleation.31 Characteristics such as contact angle, defects, and scratches influence the nucleation capability of a surface. Thus, at the same pressure, dropwise condensation can occur at slightly higher temperature than predicted by the theoretical modified equation. It was noted that in rough regions along the edges of the substrate, condensation began at a higher temperature than that shown in Figure 2. Clearly, the microcavities or edges were sites of much lowered surface energy barrier, which resulted in this phenomenon. Once the pumpdown was performed, condensation can be initiated by manipulating chamber pressure and peltier temperature, which were also the controlling parameters. This condensation process was recorded at different magnifications (250, 500, 1500, 3500, and 6500) and for an accelerating voltage of 15 keV. The observation of sub-microscopic droplets requires higher beam voltage for sharper contrast, thus the beam voltage was increased to 25 keV for better visualization. Because the removal of condensing droplets cannot be imposed inside the experimental setup, condensation, once triggered, eventually leads to complete surface coverage of the substrate. An area was selected randomly to observe a droplet growth. After the droplet merged with another droplet, the focus of electron gun was moved to another area to observe another droplet that had a higher probability of growing without merger because of lesser population density of droplets around it. With this method, droplet growth histories were recorded for different droplets at different stages of their growth. The choice of electron beam voltage was based on determination of best contrast for visualization, while also minimizing any effect of beam heating. Hence, these experiments were carried at 15 keV. Postprocessing results involved image analysis to determine droplet growth rate, which was done using ImageJ software.56 The controlling parameters for the study were chamber pressure and the peltier cooler temperature.

Results and Discussion a. Method of Condensation. As mentioned above, the condensation process can be triggered by controlling the pressure and temperature. In particular, there are two methods to achieve condensation: Mode 1: Constant vapor pressure, varying substrate temperature. The pressure is fixed and the temperature is slowly decreased to initiate dropwise condensation Mode 2: Constant substrate temperature, varying vapor pressure. The temperature of the peltier cooler is fixed and the pressure is slowly increased until condensation process starts. We found that the droplet growth and coalescence results were affected to some extent by the method use to achieve condensation. This happened because of its influence on the time required to achieve a steady state. It was observed that temperature control was much more difficult in the first mode. This resulted in nonuniform droplet growth. On the other hand, the second mode gave a more uniform temperature profile and the time required to reach steady state was drastically reduced. In this mode, a more uniform model of coalescence and droplet growth was achieved, which is discussed in the following sections.

(56) Abramoff, M. D.; Magalh~aes, P. J.; Ram, S. J. Image processing with imageJ. Biophotonics Intl. 2004, 11(7), 36–41.

17104 DOI: 10.1021/la102642r

Figure 4. Thermocouple tip.

Figure 5. Beam heating with beam voltage for line mode and spot mode.

b. Effect of Beam Heating. Because the mechanism of imaging in ESEM involves line scan of an electron beam over the surface, the impinging electrons can cause instantaneous heating of the spot at which impingement is occurring. Although the beam current may be less and the electron beam diameter is on the of order of nanometers, this can result in a very high dosage of electrons concentrated over a very small area, hence high heat flux may ensue. If the beam heating is significant, it can affect the condensation process. To ascertain the effect of beam heating, a 100 μm junction diameter thermocouple wire (Figure 4) was inserted in the ESEM chamber and was exposed to the electron beam under vacuum conditions (pressure ∼10-6 Torr). Two modes are possible—spot mode and line scan mode—and both of these modes were chosen to evaluate the effect of heating. It was found that beam heating was dependent upon several factors such as line time, beam voltage, mode type, and spot size. Figure 5 shows the effect of beam heating with line scan mode and spot mode (for spot size 3) on thermocouple tip. In general it was found that the spot mode produced a higher amount of heating and that a larger spot size produced a larger amount of heating. A maximum of 0.5 K of heating was observed under the spot mode for the current operating conditions. Increasing the spot size in general led to increase in temperature. For imaging purposes, the map scan mode was used during the experiment that involved the electron beam being rastered across an entire chosen area. For current chosen conditions of 15 keV beam voltage and line time of 16.7 ms, it can be observed that beam heating is on the order of 0.1 K. Increasing the beam voltage to 25 keV resulted in beam heating to the order of 0.3 K. It should be noted that the current beam heating experiment was performed on a thermocouple tip that is thermally and electrically conducting in nature. On insulating materials, the effect of concentrated beam of electrons may be different from present observations. Under ESEM Langmuir 2010, 26(22), 17100–17110

Anand and Son

conditions the electron beam width is enhanced to the order of micrometers because of collision events of electrons with vapor molecules inside.57,58 Thus, the dosage of electrons reaching a spot is greatly reduced, and it can be expected that beam impingement would have negligible heating effect. c. Dropwise Growth for Single Droplets: Experimental Results. As mentioned previously, understanding evolution of a single droplet can provide us with an understanding of heat transfer through the droplet, which can then be used to evaluate over all heat transfer. The droplet growth may occur by different mechanisms depending upon physical and thermodynamic conditions such as:31 (i) Direct phase change of vapor to liquid at the drop surface. (ii) Thermal release of interphase mass conversion that can lead to Marangoni forces that trigger temperature gradients inside the droplet. This can induce local subcooling at droplet periphery so probability of vapor condensation increases along the periphery. (iii) Nucleation and growth of the droplets by means of surface diffusion. This has widely been observed for water on several substrates. Experimental observations by others have led to a generalized droplet growth model of D = ktn, where D is the droplet diameter growing over a period of time t, and n and k are empirical constants. The value of n has been reported as 1/3 or 1/4 in cases where surface diffusion is a prominent mechanism.28-31 Irrespective of the mechanism involving how a droplet evolves, the condensational growth of droplet formation is a multistage process.29 The first stage involves droplet growth of single and individual droplets wherein a number of nucleation sites may become active. This stage is marked by low surface coverage. As droplets grow in size, the inter-droplet distance decreases until the droplets meet together, after which the combined effect of surface tension and growth leads to coalescence of these droplets. This is referred to as the second stage in the growth. Beyond this, in the third stage more coalescence may occur and new droplets may be nucleated in the bare area between the droplets. The essential aim of this study was to understand the first stage, involving how an individual droplet evolves over time. However, measuring droplet growth for single droplets over an extensive period of time is significantly difficult because the time for droplet interaction leading to coalescence is very short—lasting only to few number of frames. Droplet interaction depends upon the degree of subcooling with larger subcooling initiating an increased number of nucleation sites and hence increasing the population density of droplets on the surface. This then results in significant coalescence events and a much shorter time period where a single droplet can evolve without interaction from neighboring droplets. Thus, recording a substantial period of growth devoid of the coalescence behavior is a challenge. Using a very low degree of subcooling, the number of population sites can be controlled. Further, a lower degree of subcooling also decreases the droplet growth, which can further help in observation of droplet evolution over a larger period of time. The temperature of the vapor inside the ESEM chamber is 17 °C, and the temperature at which condensation occurs is -0.675 °C. From this information and by usual convention, it can be affirmed that the degree of subcooling is 17.675 °C. However, it should be noted that the presence of the peltier cooler establishes a temperature gradient around the substrate. Around the substrate ∼0 (μm), the temperature of the vapor molecules is much closer to (57) Danilatos, G. D. Foundations of environmental scanning electron microscopy. Adv. Electron. Electron Phys. 1988, 71, 109–250. (58) Khouchaf, L.; Verstraete, J. Electron scattering by gas in the Environmental Scanning Electron Microscope (ESEM): Effects on the image quality and on the X-ray microanalysis. J. Phys. IV France 2004, 118, 237–243.

Langmuir 2010, 26(22), 17100–17110

Article

Figure 6. Condensed droplets on the Si surface (4.2 Torr) (constant substrate temperature: -0.5 °C, varying vapor pressure mode).

the actual peltier cooler. Thus, the effective degree of subcooling is much lower when the peltier cooler is active. Accurate prediction of the degree of subcooling can be made by measuring temperature of vapor just above the peltier cooler surface. Because an additional thermocouple measurement cannot be incorporated in the experimental setup, prediction of an effective degree of subcooling was not achievable. By controlling the peltier cooler temperature, we were able to locate and measure droplets growing on their own without merger for time periods in the range of 40-180 s. Figure 6, shows a series of frame containing condensed droplets at 500 magnification on the silicon surface. Each droplet was identified with the initial peripheral diameter measured as the diameter of the droplet in the first frame of a series of frames where the droplet was observed to grow over a period of time. As mentioned previously, the time between successive growths is dependent upon the area and the scan rate of the electron beam. Thus, the growth record of each droplet was obtained using this technique. Figure 7 is an indicative plot of droplet growth as a function of time for the three different droplets showed in Figure 6. The different droplets are identified based on the initial diameter. The growth curve as shown in Figure 7 was found to be characteristic of droplets evolving under conditions under study. After careful analysis, many characteristics were revealed on the nature of droplets. As can be seen from Figure 7, the droplet diameter appears to grow linearly with time. A similar linear pattern was also observed when the data was plotted in terms of droplet area (D2) with time, thereby giving an impression that D2 = kt. After careful consideration, the perplexing nature of the observed drop diameter growth pattern was resolved, when it was noted that the rate of diameter growth was different for different droplets. As can be seen the rate of growth (dD/dt) was noted to decrease as the diameter of the droplet increase. The linear nature of diametrical growth for all the droplets gives a certain ambiguity to this observation. The situation becomes clearer when the diameter was observed on a logarithmic scale as shown in Figure 8. From Figure 8, it becomes unambiguously DOI: 10.1021/la102642r

17105

Article

Figure 7. Droplet diameter growth for individual droplets occurring in same frame with different initial diameters at 4.2 Torr (Constant substrate temperature, varying vapor pressure mode).

Figure 8. Droplet growth for individual droplets.

clear that the rate of diametrical growth for smaller droplets is much higher than the rate at which larger droplets appear to grow. Moreover, the droplet diametrical growth rate appears to change as the droplet size reaches near the size of mean free path of vapor molecules in the chamber. However, to understand the mechanism behind this behavior it is essential to identify an overall growth pattern where each of the droplet growth patterns can be fitted in. It is clear that droplet evolution observed for a period of 40-100 s was not enough to predict the mechanism by observing how one of the many single droplets was growing. For this purpose, a scheme was employed to visualize droplets observed at different times and different stages of growth as a single droplet growth curve. From the records of droplet growth, the droplet with smallest initial diameter (D01) analyzed was identified. Using polynomial fitting, a fourth-order equation was fitted into this curve to obtain a relation between diameter and time. Then the droplet with next smallest initial diameter (D02) was chosen to fit into the polynomial curve of the earlier droplet for identifying the time t02 at which D02 would coincide. After identifying t02, the entire series of D02 was incremented in time by t02 to obtain the next curve. This process was repeated to 17106 DOI: 10.1021/la102642r

Anand and Son

Figure 9. Droplet growth for individual droplets evaluated for overall growth.

interpolatively fit each individual droplet into the previous curve. For much larger droplets, where a pre-existing curve was not found to fit their diameters, an extrapolative scheme was used to identify the time increment with an understanding that this extrapolation scheme may not be as accurate as the interpolative scheme. Using this scheme, an overall droplet growth pattern could be observed as shown in Figure 9. From Figure 9, it can be seen that the interpolative scheme provides a good fit to an overall trend, while for larger diameters; the time increment provided by the extrapolation method may not be as accurate. These results can be improved by obtaining droplets growing in the intermittent size range; however, even in our record of over 100 single droplets with substantial growth, the droplets in between these size ranges were missing. Larger droplets increase the probability of coalescence, and when coalescence happens there is a sudden jump in droplet growth due to intermediate droplets that could not be isolated. From Figures 8 and 9, it can be seen that under the experimental conditions a droplet took a prolonged time (∼90 s) to grow from 1 to 25 μm. The behavior of droplet diameter growth provides an insight into the vapor-phase transformation, which can be understood by application of principles of kinetic theory of gases. d. Dropwise Growth for Single Droplets - Theory. The key to understanding the behavior of the observed droplet diameter growth is to identify the mechanics of vapor-liquid transformation. For steam-water transformation the most commonly occurring mechanism of this transformation involves diffusional growth. This conjecture is supported by high-magnification images (such as shown in Figure 14) that show the presence of nanosized droplets along the periphery of the growing microsized droplets. These were observed to merge into the larger droplet. Thus, diffusion of nanodroplets due to capillary imbibition is clearly active during the process of dropwise growth. The rate of growth if this mode alone is active is given by D = kt0.33.28-31 However, a large discrepancy was found when the recorded data from present studies was tried to fit with this growth law. It became clear that diffusional growth cannot solely account for the droplet growth pattern observed in the present case. The kinetics of vapor-liquid phase change behavior observed under present experimental studies can be explained by application of principles of the kinetic theory of gases. The following analysis is based on dropwise condensation theory developed by Langmuir 2010, 26(22), 17100–17110

Anand and Son

Article

condition at the interface. To analyze this problem, the evaporative component can be given by Boltzmann law.60   λs ð7Þ m_ }l ¼ m_ 0}fs exp RTv The above two equations can be solved in conjunction with heat transfer accompanying the phase change. If heat transfer is considered between vapor, condensing liquid, and the substrate, then qd ¼ he AðTv - Ti Þ ¼ hd BðTi - Tl Þ Figure 10. Droplet on a surface.

The interfacial heat transfer coefficient is related to the mass transfer occurring at the liquid-vapor interface and is given by

Umur.12 In the following analysis, the effects of disjoining pressure and extended meniscus around the droplet are neglected. Also in the accompanying theory, the nucleation characteristics of the surface are not discussed, and it is assumed that the droplet has already nucleated. A single droplet sitting on a surface at temperature Tl is considered and surrounded by vapor at temperature Tv and pressure at Pv (Figure 10). The vapor-liquid interface temperature is assumed to be Ti. The volume and the exposed surface area of the droplet sitting on the surface with contact angle θ is given by πR3 ð2 þ cos θÞð1 - cos θÞ2 V ¼ 3

ð2Þ

A ¼ 2πR2 ð1 - cos θÞ

ð3Þ

he ¼

Dbase hd 8 ¼ θ kl

Z

D=2

x dx 2

ðD=2Þ

0



9:2 ¼ f ðθÞ θ

ð10Þ

where kl ðTv - Tl Þ f1 ðθÞFl λs

ð4Þ

ð5Þ

ð6Þ

The most significant aspect of the above equation is the condensation accommodation coefficient. The condensation coefficient (Rc) defines the probability of a water molecule to stick to the condensing surface or, in other words, become a part of the condensate. Thus, (1 - Rc) signifies the molecules that are reflected back to the vapor state. Substantially lower values of the condensation coefficient signify much larger interfacial resistance and a larger temperature jump.59 Because it is assumed that the vapor-liquid interface is at a temperature Ti, it implies inclusion of a temperature jump (59) Rose, J. W. Condensation heat transfer fundamentals. Chem. Eng. Res. Des. 1998, 76(2), 143–152.

Langmuir 2010, 26(22), 17100–17110

- x2

   R - R R2 - R20 þ f2 ðθÞf ðsÞ R - R0 þ Rln ¼ η1 t R 0 - R

f1 ðθÞ ¼

ð2 þ cos θÞð1 - cos θÞ2 sin θf ðθÞ

f2 ðθÞ ¼

f ðsÞ ¼

2 - Rc 2Rc

ð11Þ

ð1 þ cos θÞf ðθÞ 2sin θ pffiffiffiffiffiffi 2πkl ðR=MÞ1:5 Tv2:5 pv λ2s

The condensation term can be given by Rc pv m_ }v ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2πTv R=M

ð9Þ

The above equations can be combined to eventually lead to

η1 ¼

At the droplet interface, impinging vapor molecules become a part of the droplet, whereas some liquid molecules escape into the vapor region. The imbalance between influx and outflux of water molecules gives the net rate of droplet growth or dissipation. In particular, the net mass flux rate at a surface is given by _ ¼ m_ }v - m_ }l m}

_ s m}λ Tv - Ti

The drop side coefficient hd can further be calculated using the drop distribution function by Mccormick and Baer.17

and the area of the droplet base is given by B ¼ πR2 sin2 θ

ð8Þ

f ðθÞ ¼

9:2 θ

The above solution is a mixed solution of linear and exponential terms. As mentioned previously, a very critical parameter in determining the solution for the above equation is the condensation accommodation coefficient Rc. Various methods employed in literature give a wide range of suitable values of Rc.61 Typically, it is expected that in pure vapor conditions Rc is closer to 1, whereas at lower pressures its value is expected to be lesser than 1. Evaluating the condensation coefficient during condensation experiments is a challenge, owing to the presence of noncondensable (60) Knacke, O.; Stranski, I. N., The mechanism of evaporation. Prog. Metal Phys. 1956, 6, (C). (61) Marek, R.; Straub, J. Analysis of the evaporation coefficient and the condensation coefficient of water. Int. J. Heat Mass Transfer 2001, 44(1), 39–53.

DOI: 10.1021/la102642r

17107

Article

Anand and Son

Figure 12. Condensing droplet at 4.2 Torr visualized at 80° inFigure 11. Theoretical and experimental droplet growth analysis.

gases and contaminations. However, as explained previously, the rarefied environmental conditions inside ESEM are conducive for such measurements. The experimental technique along with the presented theory can be used to evaluate the condensation coefficient. Although the current setup and the method can be used to determine condensation accommodation coefficient, there remains ambiguity regarding the true value on account of uncertainty in the temperature measurement. By adjusting Rc and degree of subcooling (Tv - Tl), a fit was obtained to match the theoretical results with the experimental values (Figure 11). The fit was obtained for an accommodation coefficient of 0.8 and a degree of subcooling of 0.006 °C. As has been stated before, the effective degree of subcooling for the present setup cannot be measured experimentally. Equation 10 illustrates the growth rate does not have a direct dependence upon the temperature of the substrate. Instead, the growth rate is affected by the degree of subcooling given by the temperature difference of the vapor and the substrate. Thus, the diametrical growth rate is not sensitive to the experimental temperature uncertainties of the substrate temperature. The much lower value of the effective degree of subcooling for the theoretical fit reflects the conjecture that only the molecules near the substrate contribute to the phase change process. The molecules far away from the peltier cooler are much more energetic than the molecules near the substrate surface and hence do not play an active role in the condensation process. This further indicates that the degree of the temperature jump near the liquid-vapor interface is negligible, thus the interfacial resistance is very small. Moreover, the temperature conditions inside the droplet are nearly isothermal with negligible temperature difference between the substrate surface (Ts) and the interfacial temperature (Ti) at the droplet perimeter. From the theoretical principles leading to conformity with experimental observations, it becomes clear that under the conditions of ESEM, the droplet growth occurs directly by a mechanism of conversion of vapor to the liquid medium. From Figure 11 it can be seen that droplet diameter growth rate observed during the experiment evidently changes as the droplet diameter approaches the mean free path of vapor molecules. However, this observed relevance of mean free path of vapor molecules on droplet growth is not captured in the dropwise condensation theory by Umur and requires further investigation. 17108 DOI: 10.1021/la102642r

clined surface (constant substrate temperature, varying vapor pressure mode).

Figure 13. Droplet growth rate at 4.2 Torr (constant substrate temperature, varying vapor pressure mode).

e. Dropwise Growth for Coalescing Droplets. The specially built copper stage was used to observe the coalescence behavior and the change in contact angle. Figure 12 shows an example of the droplet growth captured on the Silicon surface attached to the copper stage. Figure 13 shows the results of droplet diameter growth rate at 4.2 Torr wherein substrate temperature was held steady and vapor pressure was increased to initiate condensation. An average diameter was computed based on visible droplets in a given frame as per the equation given below. ðDavg Þframe ¼

n X

Di =n

ð12Þ

1

The diameter and contact angle of each droplet was calculated using LBADSA plugin in ImageJ.62 As seen, there was a sudden jump in droplet diameter when coalescence took place. Depending upon the number of coalescing events, the jump could be much higher. It was noticeable that after coalescence the droplets continued to grow at the same diametrical rate law as that of single droplet growth until the next coalescence took place. Another distinct pattern observable under present experimental situation was the birth of a new generation (62) Stalder, A. F.; Kulik, G.; Sage, D.; Barbieri, L.; Hoffmann, P. A snakebased approach to accurate determination of both contact points and contact angles. Colloids Surf., A 2006, 286(1-3), 92–103.

Langmuir 2010, 26(22), 17100–17110

Anand and Son

Article

Figure 14. Image Visualization at 6500 between visible droplets.

of droplets between coalescence times. Even when the drops had been nucleated and were growing, new droplets nucleated at different period of times. A substantial decrease in the average diameter (Davg)frame is indicative of a large number of new droplets that nucleate in a given frame. This is clear from Figure 13, where new droplet formation shows a decrease in average diameter, and a coalescence event led to increase in average diameter in an image frame. By observing the condensation process on the specially built copper stage, it was observed that coalescence and the droplet growth were accompanied by a change in dynamic contact angle. Particularly during the coalescence, the contact angle change depends upon the volume of merging droplets and degree of their separation. Surface forces acting on the droplet are active during all the time, and as such when droplets merge there exists a brief period during which the combined droplet tries to adjust its center of mass to satisfy those forces by overcoming frictional forces. As noted, the droplet edges are pinned down due to frictional forces due to which the combined droplet requires some time to return to its normal state.63 f. Sub-Microscopic Droplet Visualization. Droplets were found to nucleate at the same spot, which reinforces the idea that condensation is a heterogeneous nucleation phenomenon. As can be seen from Figure 14, sub-microscopic droplets were found to be nucleating and growing in between larger droplets for our case. As mentioned previously, most of the present models on dropwise condensation are based on the premise that condensation occurs primarily on the droplets with little or no condensation on portions of the surface between droplets. However, it is clear from Figure 14 that this assumption needs a revisit. Whether the formation of sub-microscopic droplets occurs everywhere at the very beginning of the first wave of droplet formation cannot be postulated yet. However, it can be seen that the “bare area” is populated entirely by sub-microscopic droplets of different sizes. Even further, from Figure 14, the presence of sub-microscopic droplets appearing, growing, and then coalescing with the larger droplet can be observed. Some of the nanodroplets can be seen sitting ∼100 nm from the larger droplet. The role of surface tension appears certainly an important one to cause diffusion/ migration of micro/nano droplets into the bigger droplet.

Conclusion Dropwise condensation was studied in a superheated environment and rarefied vapor conditions as low as 4 Torr using an ESEM. Droplet growth of single droplets was studied along with their coalescence behavior with other droplets. The analysis of (63) Leach, R. N.; Stevens, F.; Langford, S. C.; Dickinson, J. T. Dropwise condensation: Experiments and simulations of nucleation and growth of water drops in a cooling system. Langmuir 2006, 22(21), 8864–8872.

Langmuir 2010, 26(22), 17100–17110

experimental observation reveals a mechanism where growth of droplets occurs directly by means of vapor-to-liquid conversion on the droplet itself. Numerical simulations on droplet growth using a theoretical model based on the kinetic theory of gases provide a deeper understanding of interfacial mass transport under the experimental conditions. In particular, the diameter of the droplet growth was found to decrease as the droplet diameter increases. Evidence was also found that establishes the presence of sub-microscopic droplets nucleating and growing between microscopic droplets for the partially wet case. Acknowledgment. The study presented was supported by Grant U01ES016123 from the National Institute of Health (NIH). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Environmental Health Sciences or the NIH. We thank Dr. Doug Kohls and Mr. Ratandeep Singh Kukreja of the Advanced Materials Characterization Center (AMCC) at the University of Cincinnati for their invaluable help in ESEM measurements. Supporting Information Available: The video recording of droplet growth on the bare Si surface at magnifications of 250, 500, 6500, and 12 000. This material is available free of charge via the Internet at http://pubs.acs.org.

Glossary A B D D0i fs hd he kl m_ 00l m_ 00v M m_ 000 pA Psat Pw qd R R

Area of liquid-vapor interface [m2] Area of droplet base [m2] Diameter of the droplet (=2R) [m] Diameter of the droplet in the ith frame [m] Fraction of surface covered by adsorbed molecules Drop-side heat transfer coefficient [W/(m2K)] Interfacial heat transfer coefficient [W/(m2K)] Thermal conductivity of liquid [W/(K 3 m)] Rate of evaporative flux from the droplet [kg/(m2 3 s)] Rate of condensation flux toward the droplet [kg/(m2 3 s)] Molecular weight of the liquid [kg/mol] Proportionality constant Partial pressure of vapor in the gas surrounding the droplet [Pa] Saturation vapor pressure [Pa] Modified saturation vapor pressure[Pa] Rate of heat transfer through droplet [W] Radius of droplet defined by Young’s laplace equation [m] Universal gas constant [J/(mol•K)] DOI: 10.1021/la102642r

17109

Article

R0 R* t t0i Ti Tl Ts Tv V

Anand and Son

Radius of the droplet at the initial moment of observation [m] Critical radius of the droplet [m] Time [s] Elapsed time for the ith frame [second] Temperature of liquid-vapor interface [K] Temperature of liquid [K] Temperature of sample [K] Temperature of vapor [K] Volume of droplet [m3]

Greek Symbols Rc Fl λs θ

Figure A.1. Droplet on a surface.

Condensation accommodation coefficient Density of liquid [kg/m3] Latent heat of vaporization [kJ/kg] Static contact angle

Appendix: Geometrical Description of Droplet Image analysis of droplets such as shown in Figure 6 provides measurement of the diameter in terms of its base diameter (Figure A.1). If Rbase refers to the base radius of the droplet, and R refers to the radius of curvature of the droplet, then a relation between these radii and their corresponding diameters can be expressed as D ¼ 2R and Dbase ¼ 2Rbase

ðA.1Þ

As can be observed from Figure A.1, the base diameter (Dbase) of a droplet sitting on a substrate is related to the radius of curvature of the droplet (R). This relation can be expressed as R ¼ Rbase =sin θ w D ¼ Dbase =sin θ

ðA.2Þ

Since the analyzed images give the estimated size of base diameter, eq A.2 was used to convert their representation in terms of base diameter to the radius of curvature. Thus, the modified values of experimentally measured radii can be compared with the

17110 DOI: 10.1021/la102642r

theoretical predictions of eq 10, which is expressed in terms of the radius of curvature itself. The height of the droplet sitting on a surface can be related to the radius of curvature of the droplet and the contact angle the droplet is making with the surface. This relation can be given by H ¼ Rð1 - cos θÞ

ðA.3Þ

Thus, the base area of the droplet, which is the substrate area covered by the droplet, can be given by B ¼ πR2base ¼ πR2 sin2 θ

ðA.4Þ

Since a droplet sitting on a surface is essentially a part of a sphere, the interfacial area and the volume of the droplet can be expressed in terms of the radius of curvature and the contact angle as A ¼ 2πRH w A ¼ 2πR2 ð1 - cos θÞ

ðA.5Þ

and V ¼

πH 2 ð3R - HÞ πR3 ð2 þ cos θÞð1 - cos θÞ2 wV ¼ 3 3 ðA.6Þ

Langmuir 2010, 26(22), 17100–17110